Autism Spectrum Disorders (ASD) are a phenotypically heterogeneous group of neurodevelopmental disorders being diagnosed at growing rates in toddlers and young children, with prevalence estimates of 1%. Many children have limited expressive language at the time of diagnosis, and only 50% make significant gains in communication, even with intensive interventions. Unfortunately, intervention studies in ASD have struggled to reliably predict language gain, and, importantly, to differentiate pre-verbal from non-verbal children. One major limitation is the reliance on standardized behavioral assessments, which only capture overt behaviors. Innovative methodology is needed to better characterize the neural correlates of cognitive and perceptual domains important for language acquisition and to apply these methods to inform intervention studies. Such knowledge will allow us to design and apply more effective, targeted interventions for children with ASD based on baseline phenotypes. This study proposes to apply electrophysiological measures to better characterize the neural correlates of two cognitive domains essential for language acquisition: joint attention and statistical learning, along with resting state EEG power, in young children with ASD prior to entry into an intensive behavioral intervention and then after completion of the intervention, in order to determine if specific electrophysiological profiles can predict language acquisition and gain. Combined with the activities outlined in the training plan, this research proposal represents a timely effort to apply translational methods to behavioral intervention studies in order to enrich the conventional predictors of outcome in ASD. The principle unifying this study is that a better understanding of the neural mechanisms underlying the behavioral and cognitive deficits in ASD will inform predictors of treatment response, which, in turn, will facilitate the development of more targeted interventions for these children. This study is in line with objective 1 (strategy 1.3) and objective 3 (strategy 3.1) of the NIMH strategic plan. Specifically, to (1) "identify and integrate biological markers (biomarkers) and behavioral indicators associated with mental disorders," and to (2) "further develop innovative interventions and designs for intervention studies." ! PUBLIC HEALTH RELEVANCE: Language impairment is very common in children with Autism Spectrum Disorders (ASD), with 50% remaining non-verbal even with intensive treatments. Clinicians struggle to predict which children will actually make gains in language at the time of a child's diagnosis, largely because of the limitations of standardized testing to characterize these children. This study proposes to couple behavioral testing with advanced electrophysiology to predict language outcomes after treatment in children with ASD. This study has the potential to improve our ability to predict language outcomes which, in turn, will allow clinicians to make more informed treatment recommendations and to provide more detailed prognostic information to patients and their families.